In the earliest days of the Classical Prague School, Jakobson and Trubetzkoy held essentially the same views on phonological oppositions which they divided into ‘correlations’ and ‘disjunctions’. The former are binary oppositions which are shared by more than one pair of phonemes, while the latter include all other kinds of oppositions obtaining in a given language. Although Jakobson abandoned the dichotomy in his published works after 1931, Trubetzkoy retained it until his article of 1936, when he explicitly rejected the disjunction as too simplistic to account accurately for the more complex oppositions in the languages which he had studied. The gradual process whereby Trubetzkoy discarded the disjunction in favor of the more precise kinds of oppositions expounded in Grundzüge has remained unclear to historians of linguistics. The recent publication of Trubetzkoy’s letters, although somewhat disappointing on this matter, makes possible a more accurate reconstruction of Trubetzkoy’s theory of phonological oppositions.

Examination of the relevant letters suggests that Trubetzkoy had begun considering the problems associated with the disjunction no earlier than May, 1933, and no later than the end of the summer of the same year. The letters also suggest that by Nov. 1935, in an outline for his 1936 article, Trubetzkoy had finally rejected the disjunction and had arrived at a new classification of phonological oppositions, which is basically retained with a few terminological changes in Grundzüge (1939). Unfortunately, the letters offer no evidence on the reasons for Trubetzkoy’s abandonment of the disjunction beyond the explanations in his published works (1936; 1939) which are reviewed in this article. Finally, the letters reveal that the final split between Jakobson, with his theory of strictly binary oppositions, and Trubetzkoy, with his theory of both binary and non-binary oppositions, occurred soon after Trubetzkoy (1936) and not after Jakobson (1938)as Vachek (1968) has suggested.

As recent publications in phonological theory have shown, the question as to the difference between Jakobson’s and Trubetzkoy’s theories is by no means of purely historical interest. While Jakobson’s binarism and the distinctive feature approach have greatly influenced the Chomsky-Halle School and generative phonology, there has recently been a renewal of interest in certain aspects of ‘Trubetzkoyan’ phonology. The most important proponent of this new trend is P. Ladefoged who, like Trubetzkoy, bases his arguments on empirical grounds and allows multivalued as well as binary features. A few references to Ladefoged’s extensive publications reveal interesting parallels with Trubetzkoy’s views and show that the issue of strictly binary vs. multivalued oppositions is not yet resolved in modern phonological theory.